Transcript In the behavioral sciences we conduct research in order to

Analysis of Results
A lecture delivered to
participants
In
Research Methodology And Use of Technologies
In Research
On 16th Nov. (G.V.M. College of Edu.Sonepat)
Prof. Rajbir Singh
Dept. of Psychology
M.D.U.Rohtak
Why Research ?
 In
the behavioral sciences we conduct
research in order to determine the
acceptability of hypotheses which we derive
from our theories of behavior. Having
selected a certain hypothesis which seems
important in a certain theory, we collect
empirical data which should yield direct
information on the acceptability of that
hypothesis. Our decision about the
meaning of the data may lead us to retain,
revise, or reject the hypothesis and the
theory which was its source.
Theory
Hypothesis
Data
Analysis
Exhibit: Research as a process (Bidirectional)
Frequently Asked Questions before analysis
1.When analysis?
When having data/information and to draw inference
2.What type of data?
Qualitative/Quantitative?
3.If qualitative what type?
Narratives
Verbal
Product Non-Verbal
Performance
4.If quantitative what is the Level of measurement ?
Nominal, Ordinal, Interval and Ratio Scale
5. What Kind of Sample?
Large/Small:n ...N=30,----N
6.How the sample has been drawn?
Probability
Non-probability
7. Estimation of population parameters? µ, σ
8. Verification of hypothesis, if any ?
9. Correlational/ Experimental ?
10.Parametric/Nonparametric ?
11.Univariate/bivariate/multivariate ?
1, .
Results

Analysis of Data
A. Descriptive Presentation
n=1, Xs-listing, ordering, bunching, categorization
X¯, Med., Percentile, Quartile etc.
Mode: Uni/bimodal/Multimodal
Distribution- frequencies, class intervals
SD, Range,SEM
Graphic presentation- Bar diagrams, pie-charts etc.
Histogram/polygon
Data transformation- monotonic , uni-directional,calculative
destress:
√X, 1/ X,, log transformation, Arcsine transformation,
X+…, X-… etc.

Normalization T-scaling
B. Pre-Verification Test




Test of deviation from normality-SK ,Ku
Test of homogeneity- Bartlett's test, Cochran's
test
Test of homosedacity –
Range restriction-Comparing distribution
Data scanning for assumptions
C. Verification of Hypothesis/ Goodness of fit
Statistical test yields a value that has associated probability alpha, the
level of significance ( the p of committing type I error, i.e.; rejecting Ho
when it is true). Beta , the p of type II error, accepting Ho when it is
infact false. Alpha is inversely related to beta, so to reduce these errors,
we must increase N.
Sample mean and SD should be equal to population parameters. Are they
or are not? Sampling distribution of various statistic has p of x¯ and SD
to approximate of µ and σ. It is the process of estimation.

Applying a statistical test
D.Post-hoc tests: individual comparisons, Range
statistics, sample effects e.g.; Duncan’s test ,
Newman Keul’s test.
E. Interpretation: Making a Statement
Type –IV errors
Some common Tests for verification of Hypothesis
Non-parametric Test
•
Sign test, Wilcoxen
Parametric
X2
•
Median test
r
•
Mann-Whitney-U test
R
•
Kruskal -Wallis-ANOVA
•
Friedman’s ANOVA
•
Kendall’s coefficient

Spearman’s rank correlation
β
univariate
t
F
Multivariate
Manova
Canonical
Cluster
Discriminate Functions
Factor Analysis
Structural Equation
Modeling